http://rpubs.com/alex_istrate/537408

1 Analyses

2 Abstract En

2.0.1 Introduction:

In Romania, hepatitis B (HB) vaccination was introduced in the national immunization programme in November 1995 in all newborns. Two catch-up vaccination campaigns targeted unvaccinated 9 years old children between 1999 and 2008 (born between 1990 and 1998) and 18 years old adolescents between 2004 and 2008 (born between 1986 and 1990), resulting in 4 unique birth cohorts [Lexis diagram].

Yet, a higher than expected number of vaccinated young people do not have adequate protection against HB. Our objective was to assess the anti-AgHBs titres in each birth cohort.

2.0.2 Methods:

We included all patients from whom anti-AgHBs titres were measured in the Teaching Hospital of Infectious Disease in Cluj-Napoca, Romania, between March 2014 and December 2018. Exclusion criteria were: repeated examinations, confirmed history of hepatitis B during the study period and infants. We compared the anti-AgHBs titres in all birth cohorts using the Lexis surface visualization (protective titres by age, time period and cohort patterns). ~~We also evaluated the number of acute cases of hepatitis B in the same period and age groups. ~~ We also evaluated the number of acute cases of hepatitis B in the corresponding inpatient birth cohorts.

2.0.3 Results:

The final database included 2968 participants, with mean age = 31.0 ±14.2, 64.0% women. Most participants were evaluated in an ambulatory setting (78.5%).

We found that the birth-cohort (1998, 2006], vaccinated at birth (n=330, 3-dose HB vaccine coverage >90%), had a significantly lower proportion of protective titres (>10 PEI/L) - 31.2% compared to other birth cohorts vaccinated at birth: >2006 (n=106) - 67.0% or vaccinated later: (1990,1995] (n=941) - 74.5%, (1986,1990] (n=481) - 74.8%. In the unvaccinated cohort (some may have been independently vaccinated, n=1105, mean age = 45.5 ±12.4) protective titres were found in 44.7%, probably after self-limited HB infection.

Concordant results were found using median titres (PEI/L): (1998, 2006] - 3, >2006 - 29.8, (1990,1995] - 53, (1986,1990] - 108, <1986 - 4.

Despite low protective levels in the birth cohort (1998, 2006], no acute hepatitis B cases were identified during the study period in the same age group. Among the hospitalized patients corresponding to all the other vaccinated cohorts, regardless of timing, only 4 acute cases were found while and 35 cases occurred in the older age group. Despite low protective levels in the birth cohort (1998, 2006], no acute hepatitis B cases were identified during the study period in the corresponding inpatient cohort. Only 4 and 35 acute HB cases were found in the inpatient population corresponding to all the other vaccinated cohorts and the unvaccinated cohort, respectively.

Available data on a few tested infants (n=47, not included the main study), demonstrated good protection levels: 89.4%.

2.0.4 Conclusion:

We found that a sub-cohort vaccinated at birth had lower protection against hepatitis B with no clear explanation. Sporadically, acute cases were found in vaccinated cohorts suggesting that the vaccine is capable of adequate immunologic memory. However, very few cases occured in the vaccinated age groups, suggesting that the vaccine is capable of adequate immunologic memory.

The susceptible population of vaccinated individuals is 1235 (titre < 10) or 1473 (titre < 3). There were 4 acute cases corresponding to this population, resulting in observed incidences of 3.239‰ or 2.716‰. What is the incidence of acute HB in the general unvaccinated and vaccinated populations of the same ages in Romania (or good approximations)? Is our incidence higher of ~3‰ higher (meaning worse protection) or lower (meaning that the vaccine has long-lasting memory effects even at low titres) than the expected incidence in the general population?

3 Abstract Ro

3.0.1 Introducere:

Vaccinul împotriva hepatitei B (HB) a fost introdus în Programul Național de Imunizări în 1995, administrat tuturor nou-născuților. Două campanii ulterioare de vaccinare au vizat copiii de 9 ani nevaccinați (între 1999 și 2008) și adolescenții de 18 ani (între 2004 și 2008).

Există dovezi empirice că unii pacienți vaccinați nu au protecție adecvată împotriva HP. Obiectivul studiului a fost să evalueze titrurile anti-AgHBs în aceste cohorte.

3.0.2 Metode:

Am inclus toți pacienții la care au fost măsurate titrurile anti-AgHBs la Spitalul Clinic de Boli Infecțioase din Cluj-Napoca, România, între 2014 și 2018. Am exclus toate examinările repetate, pacienții cu antecedente confirmate de hepatită B și copii de 0 și 1 ani împliniți. Am comparat titrurile anti-AgHBs în toate cohortele de naștere utilizând metoda grafică Lexis (titrul protector dependent de vârstă, perioada de timp și cohortele rezultate). Am evaluat și numărul de cazuri acute de hepatită B la pacienții internați în aceeași perioadă și aparținând acelorași grupe de vârstă.

3.0.3 Rezultate:

Baza de date finală a inclus 2968 de participanți, cu vârsta medie = 31.0 ± 14.2, 64.0% de sex feminin, majoritatea examinați în spitalul de zi (78.5%).

Cohorta născută în (1998, 2006], vaccinată la naștere (n=330, acoperire vaccinală cu 3 doze >90%), a avut o proporție semnificativ mai mică de titruri protective (>10 PEI/L) - 31.2% în comparație cu alte cohorte vaccinate la naștere: >2006 (n=106) - 67.0% sau vaccinate mai târziu: (1990, 1995] (n=941) - 74.5%, (1986, 1990] (n=481) - 74.8%. În cohorta nevaccinată (n=1105, vârsta medie = 45.5 ±12.4, unii posibil vaccinați independent) am măsurat titruri protectoarela 44.7%.

Am obținut rezultate concordante și folosind titruri mediane (PEI/L): (1998, 2006] - 3, >2006 - 29.8, (1990, 1995] - 53, (1986, 1990] - 108, <1986 - 4.

În ciuda nivelurilor scăzute de protecție în cohorta (1998, 2006], nu au fost identificate cazuri de hepatită B acută internate în perioada de studiu. Între pacienții internați cu aceleași vârste ca și celelalte cohorte vaccinate, indiferent de momentul vaccinării, au fost identificate doar 4 cazuri acute și 35 de cazuri între pacienții internați născuți înainte de 1986.

Datele de la copiii de până la 1 an inclusiv (n=47, neincluși în studiul principal) au demonstrat un nivel bun de protecție: 89.4%.

3.0.4 Concluzie:

Am constatat că o sub-cohortă vaccinată la naștere avea o protecție mai slabă împotriva hepatitei B, fără nicio explicație clară. Am identificat puține cazuri acute în cohortele vaccinate, sugerând că vaccinul induce memorie imunologică adecvată.

4 Results (all data)

4.1 Descriptive table

Table 1: Summary of demographic parameters.

Factor

Detalis

spital de zi

spitalizare continua

Total

Statistics

Originea datelor

2325 (78.5%)

638 (21.5%)

2963

Anul examinarii

2014

365 (15.7%)

52 (8.2%)

417 (14.1%)

V=0.16 (p<0.001)

2015

480 (20.6%)

88 (13.8%)

568 (19.2%)

2016

515 (22.2%)

116 (18.2%)

631 (21.3%)

2017

473 (20.3%)

166 (26.0%)

639 (21.6%)

2018

492 (21.2%)

216 (33.9%)

708 (23.9%)

Sex

F

1594 (68.6%)

306 (48.0%)

1900 (64.1%)

OR=2.37 [1.98, 2.83] (p<0.001)

M

731 (31.4%)

332 (52.0%)

1063 (35.9%)

Varsta

M (min:max)

25 (2:78)

34.5 (2:97)

25 (2:97)

MW: p<0.001

μ ±DS

29.31 ±12.5

37.18 ±17.7

31.00 ±14.2

Varsta2

(0,7]

54 (2.3%)

21 (3.3%)

75 (2.5%)

V=0.30 (p<0.001)

(7,18]

127 (5.5%)

62 (9.7%)

189 (6.4%)

(18,30]

1464 (63.0%)

190 (29.8%)

1654 (55.8%)

(30,40]

305 (13.1%)

122 (19.1%)

427 (14.4%)

(40,50]

163 (7.0%)

97 (15.2%)

260 (8.8%)

(50,60]

134 (5.8%)

70 (11.0%)

204 (6.9%)

(60,70]

62 (2.7%)

53 (8.3%)

115 (3.9%)

(70,80]

16 (0.7%)

16 (2.5%)

32 (1.1%)

(80,90]

0

6 (0.9%)

6 (0.2%)

(90,100]

0

1 (0.2%)

1

Cohorta

<1986 (unvaccinated / vaccinated by request)

722 (31.1%)

383 (60.0%)

1105 (37.3%)

V=0.29 (p<0.001)

(1986,1990] (vaccinated at age 18)

402 (17.3%)

79 (12.4%)

481 (16.2%)

(1990,1995] (vaccinated at age 9)

882 (37.9%)

59 (9.2%)

941 (31.8%)

(1995,2006] (vaccinated at birth / at age 9)

244 (10.5%)

86 (13.5%)

330 (11.1%)

>2006 (vaccinated at birth)

75 (3.2%)

31 (4.9%)

106 (3.6%)

Ocupatie

copil

47 (2.2%)

18 (2.8%)

65 (2.4%)

V=0.41 (p<0.001)

elev

134 (6.3%)

65 (10.2%)

199 (7.2%)

student

984 (46.4%)

47 (7.4%)

1031 (37.4%)

salariat

797 (37.6%)

301 (47.2%)

1098 (39.8%)

lucrator pe cont propriu

25 (1.2%)

22 (3.4%)

47 (1.7%)

pensionar

88 (4.2%)

119 (18.7%)

207 (7.5%)

fara ocupatie

45 (2.1%)

66 (10.3%)

111 (4.0%)

μ ±DS = Mean (standard deviation); M (min:max) = Median (min:max); MW = Mann-WhitneyTest; OR/RR = odds-ratio / risc-ratio [95% CI] and p value from Fisher test); V = Cramer V (p value from Chi² test);

Table 2: Summary of demographic parameters.

Variabila

Detalii

<1986 (unvaccinated / vaccinated by request)

(1986,1990] (vaccinated at age 18)

(1990,1995] (vaccinated at age 9)

(1995,2006] (vaccinated at birth / at age 9)

>2006 (vaccinated at birth)

Total

Teste statistice

Cohorta

1105 (37.3%)

481 (16.2%)

941 (31.8%)

330 (11.1%)

106 (3.6%)

2963

Titru protector

494 (44.7%)

360 (74.8%)

701 (74.5%)

103 (31.2%)

71 (67.0%)

1729 (58.4%)

V=0.34 (p<0.001)

Varsta

M (min:max)

43 (28:97)

27 (24:31)

24 (19:27)

19 (8:22)

5 (2:11)

25 (2:97)

Kruskal-Wallis: p<0.001

μ ±DS

45.50 ±12.4

26.53 ±1.98

23.74 ±1.28

17.78 ±3.47

5.75 ±2.37

31.00 ±14.2

μ ±DS = Media (deviația standard); M (min:max) = Mediana (min:max); MW = Test Mann-Whitney; OR/RR = odds-ratio / risc relativ [cu IC 95%] și p calculat prin testul Fisher); V = Cramer V (p calculat prin testul Chi²);

4.2 Lexis

Figure 1: Lexis diagram.

4.3 Quantitative and qualitative data (with censoring)

Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal Statistical Society. 1995; Series B, 57, 289–300. http://www.jstor.org/stable/2346101.

Fay MP, Shaw PA. Exact and Asymptotic Weighted Logrank Tests for Interval Censored Data: The interval R package. J Stat Softw. 2010 Aug;36(2):i02. PMID: 25285054; PMCID: PMC4184046.

Figure 2: Comparrison between chohorts, using both quantitative and qualitative data. Protective titres (%).

Table 3: Summary data for the cohorts. Results from both sets of data: quantitative and qualitative, quantitative only.

Qualitative and quantitative titre data:

Quantitative titre data only:

Cohort

n=2963

% protective

% detectable

Median

Median

Geometric mean

<1986 (unvaccinated / vaccinated by request)

1105

44.7%

51.3%

4.000

23.000

7.000

(1986,1990] (vaccinated at age 18)

481

74.8%

81.1%

108.000

82.500

36.200

(1990,1995] (vaccinated at age 9)

941

74.5%

84.8%

53.000

46.000

38.900

(1995,2006] (vaccinated at birth / at age 9)

330

31.2%

49.1%

3.000

7.000

5.200

>2006 (vaccinated at birth)

106

67.0%

81.1%

29.800

32.000

26.700

Table 4: Comparrison between chohorts, using both quantitative and qualitative data. Qualitative data was considered interval-censored, therefore appropriate Logrank methods were performed according to Fay MP and Shaw PA (2010). P-values adjusted by controlling for false discovery rate, according to Benjamini Y and Hochberg Y (1995).

Pairwise

<1986 (unvaccinated / vaccinated by request)

(1986,1990] (vaccinated at age 18)

(1990,1995] (vaccinated at age 9)

(1995,2006] (vaccinated at birth / at age 9)

(1986,1990] (vaccinated at age 18)

p<0.001

(1990,1995] (vaccinated at age 9)

p<0.001

p=0.003

(1995,2006] (vaccinated at birth / at age 9)

p<0.001

p<0.001

p<0.001

>2006 (vaccinated at birth)

p=0.744

p<0.001

p=0.002

p<0.001

Asymptotic Logrank k-sample test (permutation form): p<0.001

Figure 3: Comparrison between chohorts, using both quantitative and qualitative data.

Figure 4: Comparrison between chohorts, using both quantitative and qualitative data. Qualitative data was considered interval-censored.

Table 5: Comparrison between chohorts, using both quantitative and qualitative data, Chi-square test for % with protective titres by cohort. P-values adjusted by controlling for false discovery rate, according to Benjamini Y and Hochberg Y (1995).

Pairwise

<1986 (unvaccinated / vaccinated by request)

(1986,1990] (vaccinated at age 18)

(1990,1995] (vaccinated at age 9)

(1995,2006] (vaccinated at birth / at age 9)

(1986,1990] (vaccinated at age 18)

p<0.001

(1990,1995] (vaccinated at age 9)

p<0.001

p=0.937

(1995,2006] (vaccinated at birth / at age 9)

p<0.001

p<0.001

p<0.001

>2006 (vaccinated at birth)

p<0.001

p=0.138

p=0.138

p<0.001

Chi-squared test: p<0.001

Figure 5: Comparrison between chohorts, using both quantitative and qualitative data. Protective titres (%).

Qualitative and quantitative titre data:

Quantitative titre data only:

Cohort

n=2963

% protective

% detectable

Median

Median

Geometric mean

<1986 (D)

1105

44.7%

51.3%

4.000

23.000

7.000

(1986,1990] (C)

481

74.8%

81.1%

108.000

82.500

36.200

(1990,1995] (B2)

941

74.5%

84.8%

53.000

46.000

38.900

(1995,1998] (A2,B1)

186

34.9%

50.0%

3.100

7.000

5.700

(1998,2006] (A1b)

144

26.4%

47.9%

3.000

6.500

4.500

>2006 (A1a)

106

67.0%

81.1%

29.800

32.000

26.700

Table 6: Summary data for the cohorts. Results from both sets of data: quantitative and qualitative, quantitative only.

Qualitative and quantitative titre data:

Quantitative titre data only:

Cohort

n=2963

% protective

% detectable

Median

Median

Geometric mean

<1986 (D)

1105

44.7%

51.3%

4.000

23.000

7.000

(1986,1990] (C)

481

74.8%

81.1%

108.000

82.500

36.200

(1990,1995] (B2)

941

74.5%

84.8%

53.000

46.000

38.900

(1995,1998] (A2,B1)

186

34.9%

50.0%

3.100

7.000

5.700

(1998,2006] (A1b)

144

26.4%

47.9%

3.000

6.500

4.500

>2006 (A1a)

106

67.0%

81.1%

29.800

32.000

26.700

Table 7: Comparrison between chohorts, using both quantitative and qualitative data. Qualitative data was considered interval-censored, therefore appropriate Logrank methods were performed according to Fay MP and Shaw PA (2010). P-values adjusted by controlling for false discovery rate, according to Benjamini Y and Hochberg Y (1995).

Pairwise

<1986 (D)

(1986,1990] (C)

(1990,1995] (B2)

(1995,1998] (A2=B1)

(1998,2006] (A1b)

(1986,1990] (C)

p<0.001

(1990,1995] (B2)

p<0.001

p=0.003

(1995,1998] (A2=B1)

p<0.001

p<0.001

p<0.001

(1998,2006] (A1b)

p<0.001

p<0.001

p<0.001

p=0.265

>2006 (A1a)

p=0.744

p<0.001

p=0.002

p<0.001

p<0.001

Asymptotic Logrank k-sample test (permutation form): p<0.001

Figure 6: Comparrison between chohorts, using both quantitative and qualitative data.

Figure 7: Comparrison between chohorts, using both quantitative and qualitative data. Qualitative data was considered interval-censored.

Table 8: Comparrison between chohorts, using both quantitative and qualitative data, Chi-square test for % with protective titres by cohort. P-values adjusted by controlling for false discovery rate, according to Benjamini Y and Hochberg Y (1995).

Pairwise

<1986 (D)

(1986,1990] (C)

(1990,1995] (B2)

(1995,1998] (A2=B1)

(1998,2006] (A1b)

(1986,1990] (C)

p<0.001

(1990,1995] (B2)

p<0.001

p=0.937

(1995,1998] (A2=B1)

p=0.022

p<0.001

p<0.001

(1998,2006] (A1b)

p<0.001

p<0.001

p<0.001

p=0.133

>2006 (A1a)

p<0.001

p=0.133

p=0.133

p<0.001

p<0.001

Chi-squared test: p<0.001

4.4 Quantitative data

Figure  8: Comparrison between chohorts, violin plots using log-transformed quantitative data.

Table  9: Comparrison between chohorts, using only quantitative data. P-values adjusted by controlling for false discovery rate, according to Benjamini Y and Hochberg Y (1995).

Pairwise

<1986 (unvaccinated / vaccinated by request)

(1986,1990] (vaccinated at age 18)

(1990,1995] (vaccinated at age 9)

(1995,2006] (vaccinated at birth / at age 9)

(1986,1990] (vaccinated at age 18)

p<0.001

(1990,1995] (vaccinated at age 9)

p<0.001

p=0.011

(1995,2006] (vaccinated at birth / at age 9)

p<0.001

p<0.001

p<0.001

>2006 (vaccinated at birth)

p=0.174

p=0.011

p=0.328

p<0.001

Kruskal-Wallis rank sum test: p<0.001

Table  10: Comparrison between chohorts, using only quantitative data. P-values from Conover-Iman post-hoc test.
  Kruskal-Wallis rank sum test

data: x and group
Kruskal-Wallis chi-squared = 150.6634, df = 4, p-value = 0

                           Comparison of x by group                            
                             (Benjamini-Yekuteili)                             
Col Mean-|
Row Mean |   (1986,19   (1990,19   (1995,20   <1986 (u
---------+--------------------------------------------
(1990,19 |   2.432589
         |     0.0276
         |
(1995,20 |   10.85981   10.14514
         |    0.0000*    0.0000*
         |
<1986 (u |   7.808285   6.722609  -5.162988
         |    0.0000*    0.0000*    0.0000*
         |
>2006 (v |   2.747377   1.471862  -4.970320  -1.869618
         |    0.0127*     0.2068    0.0000*     0.1004

alpha = 0.05
Reject Ho if p <= alpha/2
Figure  9: Comparrison between chohorts, violin plots using log-transformed quantitative data.

Table  11: Comparrison between chohorts, using only quantitative data. P-values adjusted by controlling for false discovery rate, according to Benjamini Y and Hochberg Y (1995).

Pairwise

<1986 (D)

(1986,1990] (C)

(1990,1995] (B2)

(1995,1998] (A2=B1)

(1998,2006] (A1b)

(1986,1990] (C)

p<0.001

(1990,1995] (B2)

p<0.001

p=0.014

(1995,1998] (A2=B1)

p=0.014

p<0.001

p<0.001

(1998,2006] (A1b)

p=0.009

p<0.001

p<0.001

p=1.000

>2006 (A1a)

p=0.205

p=0.014

p=0.398

p<0.001

p<0.001

Kruskal-Wallis rank sum test: p<0.001

Table  12: Comparrison between chohorts, using only quantitative data. P-values from Conover-Iman post-hoc test.
  Kruskal-Wallis rank sum test

data: x and group
Kruskal-Wallis chi-squared = 151.0982, df = 5, p-value = 0

                           Comparison of x by group                            
                             (Benjamini-Yekuteili)                             
Col Mean-|
Row Mean |   (1986,19   (1990,19   (1995,19   (1998,20   <1986 (D
---------+-------------------------------------------------------
(1990,19 |   2.432236
         |     0.0313
         |
(1995,19 |   8.726587   7.776082
         |    0.0000*    0.0000*
         |
(1998,20 |   8.421705   7.504513   0.687677
         |    0.0000*    0.0000*     0.8159
         |
<1986 (D |   7.807150   6.721632  -3.801978  -4.105295
         |    0.0000*    0.0000*    0.0004*    0.0001*
         |
>2006 (A |   2.746978   1.471648  -4.241376  -4.544194  -1.869346
         |    0.0137*     0.2512    0.0001*    0.0000*     0.1182

alpha = 0.05
Reject Ho if p <= alpha/2

5 Results: ambulatory data only

5.1 Descriptive table

Table 2: Summary of demographic parameters.

Variabila

Detalii

<1986 (unvaccinated / vaccinated by request)

(1986,1990] (vaccinated at age 18)

(1990,1995] (vaccinated at age 9)

(1995,2006] (vaccinated at birth / at age 9)

>2006 (vaccinated at birth)

Total

Teste statistice

Cohorta

722 (31.1%)

402 (17.3%)

882 (37.9%)

244 (10.5%)

75 (3.2%)

2325

Titru protector

327 (45.3%)

310 (77.1%)

656 (74.4%)

82 (33.6%)

50 (66.7%)

1425 (61.3%)

V=0.34 (p<0.001)

Varsta

M (min:max)

41 (28:78)

26 (24:31)

24 (19:27)

19 (8:22)

5 (2:11)

25 (2:78)

Kruskal-Wallis: p<0.001

μ ±DS

44.10 ±11.6

26.28 ±1.96

23.71 ±1.25

17.94 ±3.59

5.83 ±2.27

29.31 ±12.5

μ ±DS = Media (deviația standard); M (min:max) = Mediana (min:max); MW = Test Mann-Whitney; OR/RR = odds-ratio / risc relativ [cu IC 95%] și p calculat prin testul Fisher); V = Cramer V (p calculat prin testul Chi²);

5.2 Lexis

Figure 1: Lexis diagram.

5.3 Quantitative and qualitative data (with censoring)

Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal Statistical Society. 1995; Series B, 57, 289–300. http://www.jstor.org/stable/2346101.

Fay MP, Shaw PA. Exact and Asymptotic Weighted Logrank Tests for Interval Censored Data: The interval R package. J Stat Softw. 2010 Aug;36(2):i02. PMID: 25285054; PMCID: PMC4184046.

Figure 2: Comparrison between chohorts, using both quantitative and qualitative data. Protective titres (%).

Table 3: Summary data for the cohorts. Results from both sets of data: quantitative and qualitative, quantitative only.

Qualitative and quantitative titre data:

Quantitative titre data only:

Cohort

n=2325

% protective

% detectable

Median

Median

Geometric mean

<1986 (unvaccinated / vaccinated by request)

722

45.3%

51.5%

5.000

22.500

6.700

(1986,1990] (vaccinated at age 18)

402

77.1%

83.6%

120.500

86.000

41.100

(1990,1995] (vaccinated at age 9)

882

74.4%

84.6%

53.000

45.500

38.700

(1995,2006] (vaccinated at birth / at age 9)

244

33.6%

51.6%

4.000

7.000

5.900

>2006 (vaccinated at birth)

75

66.7%

81.3%

27.000

31.000

27.100

Table 4: Comparrison between chohorts, using both quantitative and qualitative data. Qualitative data was considered interval-censored, therefore appropriate Logrank methods were performed according to Fay MP and Shaw PA (2010). P-values adjusted by controlling for false discovery rate, according to Benjamini Y and Hochberg Y (1995).

Pairwise

<1986 (unvaccinated / vaccinated by request)

(1986,1990] (vaccinated at age 18)

(1990,1995] (vaccinated at age 9)

(1995,2006] (vaccinated at birth / at age 9)

(1986,1990] (vaccinated at age 18)

p<0.001

(1990,1995] (vaccinated at age 9)

p<0.001

p=0.002

(1995,2006] (vaccinated at birth / at age 9)

p<0.001

p<0.001

p<0.001

>2006 (vaccinated at birth)

p=0.826

p<0.001

p=0.006

p<0.001

Asymptotic Logrank k-sample test (permutation form): p<0.001

Figure 3: Comparrison between chohorts, using both quantitative and qualitative data.

Figure 4: Comparrison between chohorts, using both quantitative and qualitative data. Qualitative data was considered interval-censored.

Figure 5: Comparrison between chohorts, using both quantitative and qualitative data. Protective titres (%).

Qualitative and quantitative titre data:

Quantitative titre data only:

Cohort

n=2325

% protective

% detectable

Median

Median

Geometric mean

<1986 (D)

722

45.3%

51.5%

5.000

22.500

6.700

(1986,1990] (C)

402

77.1%

83.6%

120.500

86.000

41.100

(1990,1995] (B2)

882

74.4%

84.6%

53.000

45.500

38.700

(1995,1998] (A2,B1)

151

34.4%

49.0%

3.000

6.000

5.900

(1998,2006] (A1b)

93

32.3%

55.9%

4.000

7.700

6.100

>2006 (A1a)

75

66.7%

81.3%

27.000

31.000

27.100

Table 6: Summary data for the cohorts. Results from both sets of data: quantitative and qualitative, quantitative only.

Qualitative and quantitative titre data:

Quantitative titre data only:

Cohort

n=2325

% protective

% detectable

Median

Median

Geometric mean

<1986 (D)

722

45.3%

51.5%

5.000

22.500

6.700

(1986,1990] (C)

402

77.1%

83.6%

120.500

86.000

41.100

(1990,1995] (B2)

882

74.4%

84.6%

53.000

45.500

38.700

(1995,1998] (A2,B1)

151

34.4%

49.0%

3.000

6.000

5.900

(1998,2006] (A1b)

93

32.3%

55.9%

4.000

7.700

6.100

>2006 (A1a)

75

66.7%

81.3%

27.000

31.000

27.100

Table 7: Comparrison between chohorts, using both quantitative and qualitative data. Qualitative data was considered interval-censored, therefore appropriate Logrank methods were performed according to Fay MP and Shaw PA (2010). P-values adjusted by controlling for false discovery rate, according to Benjamini Y and Hochberg Y (1995).

Pairwise

<1986 (D)

(1986,1990] (C)

(1990,1995] (B2)

(1995,1998] (A2=B1)

(1998,2006] (A1b)

(1986,1990] (C)

p<0.001

(1990,1995] (B2)

p<0.001

p=0.002

(1995,1998] (A2=B1)

p<0.001

p<0.001

p<0.001

(1998,2006] (A1b)

p=0.009

p<0.001

p<0.001

p=0.645

>2006 (A1a)

p=0.826

p<0.001

p=0.007

p<0.001

p=0.002

Asymptotic Logrank k-sample test (permutation form): p<0.001

Figure 6: Comparrison between chohorts, using both quantitative and qualitative data.

Figure 7: Comparrison between chohorts, using both quantitative and qualitative data. Qualitative data was considered interval-censored.

5.4 Quantitative data

Figure  8: Comparrison between chohorts, violin plots using log-transformed quantitative data.

Table  9: Comparrison between chohorts, using only quantitative data. P-values adjusted by controlling for false discovery rate, according to Benjamini Y and Hochberg Y (1995).

Pairwise

<1986 (unvaccinated / vaccinated by request)

(1986,1990] (vaccinated at age 18)

(1990,1995] (vaccinated at age 9)

(1995,2006] (vaccinated at birth / at age 9)

(1986,1990] (vaccinated at age 18)

p<0.001

(1990,1995] (vaccinated at age 9)

p<0.001

p=0.004

(1995,2006] (vaccinated at birth / at age 9)

p=0.011

p<0.001

p<0.001

>2006 (vaccinated at birth)

p=0.284

p=0.013

p=0.375

p<0.001

Kruskal-Wallis rank sum test: p<0.001

Table  10: Comparrison between chohorts, using only quantitative data. P-values from Conover-Iman post-hoc test.
  Kruskal-Wallis rank sum test

data: x and group
Kruskal-Wallis chi-squared = 129.3159, df = 4, p-value = 0

                           Comparison of x by group                            
                             (Benjamini-Yekuteili)                             
Col Mean-|
Row Mean |   (1986,19   (1990,19   (1995,20   <1986 (u
---------+--------------------------------------------
(1990,19 |   2.973582
         |    0.0063*
         |
(1995,20 |   10.10705   8.943546
         |    0.0000*    0.0000*
         |
<1986 (u |   7.826020   6.259908  -3.900130
         |    0.0000*    0.0000*    0.0002*
         |
>2006 (v |   2.870802   1.428753  -3.948454  -1.613060
         |    0.0076*     0.2245    0.0002*     0.1740

alpha = 0.05
Reject Ho if p <= alpha/2
Figure  9: Comparrison between chohorts, violin plots using log-transformed quantitative data.

Table  11: Comparrison between chohorts, using only quantitative data. P-values adjusted by controlling for false discovery rate, according to Benjamini Y and Hochberg Y (1995).

Pairwise

<1986 (D)

(1986,1990] (C)

(1990,1995] (B2)

(1995,1998] (A2=B1)

(1998,2006] (A1b)

(1986,1990] (C)

p<0.001

(1990,1995] (B2)

p<0.001

p=0.004

(1995,1998] (A2=B1)

p=0.065

p<0.001

p<0.001

(1998,2006] (A1b)

p=0.129

p<0.001

p<0.001

p=1.000

>2006 (A1a)

p=0.334

p=0.017

p=0.455

p<0.001

p=0.002

Kruskal-Wallis rank sum test: p<0.001

Table  12: Comparrison between chohorts, using only quantitative data. P-values from Conover-Iman post-hoc test.
  Kruskal-Wallis rank sum test

data: x and group
Kruskal-Wallis chi-squared = 129.3764, df = 5, p-value = 0

                           Comparison of x by group                            
                             (Benjamini-Yekuteili)                             
Col Mean-|
Row Mean |   (1986,19   (1990,19   (1995,19   (1998,20   <1986 (D
---------+-------------------------------------------------------
(1990,19 |   2.972631
         |    0.0075*
         |
(1995,19 |   8.493213   7.222063
         |    0.0000*    0.0000*
         |
(1998,20 |   7.235820   6.040669   0.257292
         |    0.0000*    0.0000*     1.0000
         |
<1986 (D |   7.823515   6.257904  -3.149658  -2.846675
         |    0.0000*    0.0000*    0.0046*    0.0093*
         |
>2006 (A |   2.869883   1.428296  -3.566964  -3.414163  -1.612544
         |    0.0094*     0.2727    0.0013*    0.0020*     0.2050

alpha = 0.05
Reject Ho if p <= alpha/2

6 Alte texte

Since the introduction of HB vaccine in 1995 in Romania, there were several nation-wide vaccination campaigns: These campaigns result in several cohorts based on vaccination policies: A1 - born after 1998, vaccinated at birth; A2, same as B1 - born in (1995, 1998], vaccinated at birth or at 9 years old; B2 - born in (1990, 1995], vaccinated at 9 years old; C - born in (1986, 1990], vaccinated at 18 years old; D - born before 1986, unvaccinated.

We found that a subgroup of the A1 cohort, born in (1998, 2006], vaccinated at birth, had significantly lower protection against hepatitis B based on a low proportion of persons with titres > 10 PEI/L (31.2%) as well as a low median titre (3 PEI/L), compared to the other vaccinated cohorts.

The same cohort had significantly lower median titre (3 PEI/L) compared to other vaccinated birth cohorts: >2006 - 29.8, (1990,1995] - 53, (1986,1990] - 108, similar to the unvaccinated cohort - 4.

Impartirea pe cohorte

Since the introduction of HB vaccine in 1991 in Romania, there were several nation-wide vaccination campaigns:

.